The discussion surrounding artificial intelligence (AI) and its impact on human employment has reached new heights. However, Stephen Chin, Vice President – Developer Relations at Neo4j, offers a more detailed perspective to the developer community: AI won’t replace your job, but a developer proficient in AI will.
“Developers who fail to adapt and learn AI are the ones at a disadvantage,” Chin shared with CNBC-TV18 during the Great International Developer Summit 2026 (GIDS 2026). “Success belongs to those willing to upskill, explore new AI technologies, and gather the necessary tools.”
Chin was blunt about the current state of the developer landscape. He highlighted a significant drop in the employment rates of computer science graduates since the arrival of ChatGPT, explaining that it’s not a matter of companies needing fewer engineers, but rather that expectations have shifted.
“Most computer science graduates followed traditional curricula and didn’t learn to utilize AI tools,” he noted. “When companies interview them and ask them to use tools like Claude Code or Cursor to tackle a problem, they often respond — ‘We weren’t permitted to use AI in school; it’s cheating.’ As a result, they miss out on job opportunities.”
His message to universities is clear: update the curriculum. Equally, he advised parents to introduce their children to AI tools at an early age.
“Provide them with a Copilot subscription and encourage them to experiment with Claude Code. If you help them become proficient in AI, they’ll gain a significant advantage. They’ll graduate as AI engineers who surpass many current senior developers.”
AI enhances human capability, rather than replacing it
Addressing the larger question of whether AI will completely replace human roles, Chin offered an insightful comparison. He explained that humans often attribute human-like traits to entities that only mimic them, which doesn’t make these entities human.
ALSO READ | AI might not eliminate white-collar jobs as quickly as expected: Raghuram Rajan
“AI excels in specific problem-solving scenarios. LLMs have access to data sets far larger than any human experiences in a lifetime. However, they falter in areas like complex reasoning, requirement comprehension, and accountability,” he remarked. “This is where humans play a crucial role. AI enhances human effort — it doesn’t substitute it.”
He also pointed out a subtle yet important observation: experienced developers working with familiar code may actually work slightly slower when using AI tools. “There’s a trade-off. Junior developers gain significant enhancements, while seasoned developers on new code bases can adapt quickly. But experienced developers who are already acclimated to a familiar code base? They are somewhat slower with AI.”
A notable transformation Chin emphasized is the shift from AI experimentation to actual implementation. Companies have transitioned from running pilots and proofs of concept that rarely reach real-world usage to having successful applications where AI drives enterprise and production systems.
“The significant change this year has been from merely experimenting with AI — where many projects failed — to successful implementations where AI is genuinely powering production systems,” he stated.
Chin noted that one major reason for early deployments’ failures was the excessive dependence on generic RAG (retrieval augmented generation) systems, which often struggle with accuracy and explainability.
More companies are now gravitating towards agentic architectures supported by knowledge graphs and graph databases, enabling AI systems to operate with a structured and relationship-aware data layer.
“Even Anthropic and several other AI companies recommend integrating knowledge graphs into system architecture,” he mentioned. “This marks the transition from research to a practical enabler for application deployment.” Neo4j counts AbbVie, Pfizer, and Daimler among its clients who have utilized this approach to expedite their AI production.
On AI costs: Stay informed, but think strategically
As Indian companies increasingly partner with AI providers like Anthropic, concerns about subscription and token costs are rising. Chin acknowledged these concerns but stated that progress is being made.
“The costs associated with LLMs are decreasing as models improve in efficiency,” he explained. “Initially, many token costs were subsidized — OpenAI has been operating at a loss for some time. However, as models become more efficient and competition heightens, token prices will decline.”
ALSO READ | AI vs jobs: AI cannot be a free pass for layoffs, signals China courts in landmark rulings
In the meantime, he recommended adopting smarter architectures. “By utilizing a better knowledge framework — like a graph database — you can deploy a smaller, more affordable model and still achieve results comparable to the latest AI models. This saves costs today, and in the future, costs will continue to decrease.”
Which industries will benefit the most
Sectors such as healthcare, finance, supply chain, customer service, and legal research are among those Chin believes will benefit significantly from AI adoption.
“In finance, success in trading or fraud detection hinges on the ability to analyze large datasets, detect patterns swiftly, and react quicker than competitors,” he noted. “An accelerated system capable of processing more data presents a substantial competitive edge.”
When discussing trust in AI for critical decisions, he remained cautious. “Humans make errors too. For mission-critical AI systems, it’s crucial to involve a human to review final transactions.” He referenced context graphs — which allow AI systems to learn from past choices — as a framework that enhances AI reliability within sensitive domains like credit assessments or fraud detection.
Chin’s advice for Indian companies aspiring to accelerate AI adoption is straightforward: close the skills gap, create the appropriate architecture, and stop viewing AI as either a magic solution or a threat.
“AI already offers numerous practical and pragmatic applications,” he stated. “These systems are here to remain as they are transforming both development and business positively — improving efficiency, productivity, and enabling more with less.”
ALSO READ | AI could contribute over $500 billion to India’s economy by 2030: Study
For developers uncertain about the future, his final message is direct: take ownership of your code, utilize AI wherever it makes sense, and never allow AI tool outputs to go unexamined. “You are the developer. You hold responsibility for the code. Treat everything — both the code you write and that produced by AI — as your work, maintaining high standards.”
